skip to main content
OSTI.GOV title logo U.S. Department of Energy
Office of Scientific and Technical Information

Title: Predicting Elastic Properties of Materials from Electronic Charge Density Using 3D Deep Convolutional Neural Networks

Journal Article · · Journal of Physical Chemistry. C
ORCiD logo [1];  [2];  [2];  [1]; ORCiD logo [1]; ORCiD logo [3]
  1. Univ. of South Carolina, Columbia, SC (United States)
  2. Univ. of South Carolina, Columbia, SC (United States); Dalian Univ. of Technology (China). Key Lab. of Ocean Energy
  3. Univ. of South Carolina, Columbia, SC (United States); Guizhou Univ., Guiyang (China)
+ Show Author Affiliations

Materials representation plays a key role in machine learning-based prediction of materials properties and new materials discovery. Currently both graph and three-dimensional (3D) voxel representation methods are based on the heterogeneous elements of the crystal structures. Here, we propose to use electronic charge density (ECD) as a generic unified 3D descriptor for materials property prediction with the advantage of possessing close relation with the physical and chemical properties of materials. We developed an ECD-based 3D convolutional neural networks (CNNs) for predicting the elastic properties of materials, in which CNNs can learn effective hierarchical features with multiple convolving and pooling operations. Extensive benchmark experiments over 2170 $$Fm\bar3m$$ face-centered-cubic materials show that our ECD-based CNNs can achieve good performance for elasticity prediction. Especially, our CNN models based on the fusion of elemental Materials-Agnostic Platform for Informatics and Exploration features and ECD descriptors achieved the best fivefold cross-validation performance. More importantly, we showed that our ECD-based CNN models can achieve significantly better extrapolation performance when evaluated over nonredundant data sets, where there are few neighbor-training samples around test samples. As an additional validation, we evaluated the predictive performance of our models on 329 materials of space group $$Fm\bar3m$$ by comparing to density functional theory calculated values, which shows a better prediction power of our model for bulk modulus than shear modulus. Because of the unified representation power of ECD, it is expected that our ECD-based CNN approach can also be applied to predict other physical and chemical properties of crystalline materials.

Research Organization:
Univ. of South Carolina, Columbia, SC (United States)
Sponsoring Organization:
USDOE Office of Science (SC); National Science Foundation (NSF)
Grant/Contract Number:
SC0020272; 1940099; 1905775; OIA-1655740
OSTI ID:
1803944
Journal Information:
Journal of Physical Chemistry. C, Vol. 124, Issue 31; ISSN 1932-7447
Publisher:
American Chemical SocietyCopyright Statement
Country of Publication:
United States
Language:
English

References (56)

Atom-density representations for machine learning journal April 2019
Classification of chemical bonds based on topological analysis of electron localization functions journal October 1994
Predicting the effective thermal conductivity of composites from cross sections images using deep learning methods journal November 2019
A Universal 3D Voxel Descriptor for Solid-State Material Informatics with Deep Convolutional Neural Networks journal December 2017
Accuracy and transferability of Gaussian approximation potential models for tungsten journal September 2014
Predicting the effective thermal conductivities of composite materials and porous media by machine learning methods journal December 2018
Convolutional Neural Networks for Crystal Material Property Prediction Using Hybrid Orbital-Field Matrix and Magpie Descriptors journal April 2019
Band Gap Prediction for Large Organic Crystal Structures with Machine Learning journal July 2019
Predicting charge density distribution of materials using a local-environment-based graph convolutional network journal November 2019
Including crystal structure attributes in machine learning models of formation energies via Voronoi tessellations journal July 2017
External electric field driving the ultra-low thermal conductivity of silicene journal January 2017
On representing chemical environments journal May 2013
Machine learning reveals orbital interaction in materials journal January 2017
Evaluating explorative prediction power of machine learning algorithms for materials discovery using k -fold forward cross-validation journal January 2020
Inverse Design of Solid-State Materials via a Continuous Representation journal November 2019
Crystal structures and elastic properties of superhard Ir N 2 and Ir N 3 from first principles journal August 2007
A fast and robust algorithm for Bader decomposition of charge density journal June 2006
Machine learning reveals orbital interaction in materials dataset January 2017
Predicting the Band Gaps of Inorganic Solids by Machine Learning journal March 2018
Machine learning for quantum mechanics in a nutshell journal July 2015
Ab initiomolecular dynamics for liquid metals journal January 1993
Lone-pair electrons induced anomalous enhancement of thermal transport in strained planar two-dimensional materials journal August 2018
Materials discovery and design using machine learning journal September 2017
Graph Networks as a Universal Machine Learning Framework for Molecules and Crystals journal April 2019
Berechnung der Seitenversetzung des totalreflektierten Strahles journal January 1948
Electronic Structure, Electronic Charge Density, and Optical Properties Analysis of GdX 3 (X = In, Sn, Tl, and Pb) Compounds: DFT Calculations journal August 2015
Can machine learning identify the next high-temperature superconductor? Examining extrapolation performance for materials discovery journal January 2018
Comparing molecules and solids across structural and alchemical space journal January 2016
A quantitative uncertainty metric controls error in neural network-driven chemical discovery journal January 2019
Fast and Accurate Modeling of Molecular Atomization Energies with Machine Learning journal January 2012
Accelerating materials property predictions using machine learning journal September 2013
Machine learning models for the lattice thermal conductivity prediction of inorganic materials journal December 2019
Machine Learning-Assisted Discovery of Solid Li-Ion Conducting Materials journal November 2018
Efficiency of ab-initio total energy calculations for metals and semiconductors using a plane-wave basis set journal July 1996
Predicting the Electrochemical Properties of Lithium-Ion Battery Electrode Materials with the Quantum Neural Network Algorithm journal February 2019
VoxNet: A 3D Convolutional Neural Network for real-time object recognition conference September 2015
Predicting superhard materials via a machine learning informed evolutionary structure search journal September 2019
Crystal structure representations for machine learning models of formation energies journal April 2015
Matminer: An open source toolkit for materials data mining journal September 2018
Efficient iterative schemes for ab initio total-energy calculations using a plane-wave basis set journal October 1996
Squeeze-and-Excitation Networks conference June 2018
Creating Machine Learning-Driven Material Recipes Based on Crystal Structure journal January 2019
Data-driven atomic environment prediction for binaries using the Mendeleev number journal March 2004
Enhancing materials property prediction by leveraging computational and experimental data using deep transfer learning journal November 2019
Multi-view 3D Object Detection Network for Autonomous Driving conference July 2017
Materials Design and Discovery with High-Throughput Density Functional Theory: The Open Quantum Materials Database (OQMD) journal September 2013
Screening billions of candidates for solid lithium-ion conductors: A transfer learning approach for small data journal June 2019
Competing mechanism driving diverse pressure dependence of thermal conductivity of X Te ( X = Hg , Cd ,   and   Zn ) journal December 2015
Big Data of Materials Science: Critical Role of the Descriptor journal March 2015
Gradient-based learning applied to document recognition journal January 1998
Material structure-property linkages using three-dimensional convolutional neural networks journal March 2018
A general-purpose machine learning framework for predicting properties of inorganic materials journal August 2016
Crystal Graph Convolutional Neural Networks for an Accurate and Interpretable Prediction of Material Properties journal April 2018
Random Forests journal January 2001
Combinatorial screening for new materials in unconstrained composition space with machine learning journal March 2014
Inception-v4, Inception-ResNet and the Impact of Residual Connections on Learning journal February 2017

Similar Records

Related Subjects

36 MATERIALS SCIENCE
Elements
Young’s modulus
Elastic materials
Layers
Materials